节点文献
少数民族语言分词技术评测数据集MLWS2021
A dataset of word segmentation technology evaluation for minority languages(MLWS2021)
【摘要】 依据蒙古文、藏文和维吾尔文词汇的构词规律和特点,制定适合计算机信息处理的蒙古文、藏文和维吾尔文分词评测标准,构建蒙古文、藏文和维吾尔文的分词标注语料,形成标准评测数据集(MLWS2021),为解决自动分词、词性标注、信息检索、语料库构建等研究课题提供依据。MLWS2021共包含2.5万句藏文、6.5万句蒙古文、6.5万句维吾尔文。本评测数据集将面向社会,提供免费评测服务,逐步建成权威的少数民族语言分词技术评测平台,推动少数民族语言信息处理技术的发展。
【Abstract】 According to the morphological rules and characteristics of Mongolian,Tibetan and Uyghur vocabulary,we formulated the evaluation standards for Mongolian,Tibetan and Uyghur word segmentation suitable for computer information processing,established the word segmentation and tagging corpus of Mongolian,Tibetan and Uyghur,and formed a standard evaluation corpus(MLWS2021),so as to solve the problems of automatic word segmentation,part of speech tagging,information retrieval.MLWS2021 contains 25,000 Tibetan sentences,65,000 Mongolian sentences and 65,000 Uighur sentences.The evaluation dataset will provide free evaluation services available to the public,gradually build an authoritative minority language evaluation platform,and promote the development of minority language information processing technology.
【Key words】 minority language; word segmentation; evaluation dataset; standard specification for word segmentation;
- 【文献出处】 中国科学数据(中英文网络版) ,China Scientific Data , 编辑部邮箱 ,2022年02期
- 【分类号】H2;TP391.1
- 【下载频次】97